Path Loss Modelization in VHF and UHF Systems

1 Path Loss Modelization in VHF and UHF Systems Tiago A. A. Rodrigues, António J. C. B. Rodrigues Abstract The main purpose of this paper is to assess the recommendation ITU-R P.46-3 proposed by the International Communications Union resulting of an evolution from the original version, ITU-R P.46. The method present here is based on the difference in calculations between measured values and simulated values, respecting this recommendation, provided by the MONITORplus program. The measured values were taken from four different environments, free space, quasi open, forest and sub-urban in mainland Portugal. Index Terms Free space, forest, propagation, quasi open recommendation ITU-R P.46-3, sub-urban. T I. INTRODUCTION HE mobile wireless and private networks had a huge boom in the past 2 decades. Since telecommunications are a source of progress and economic development and knowing that radio-electric spectrum is a scarce resource, there must be a correct management of it, monitoring and optimization by an authority. It is in this environment that the present paper is inserted [1][7]. The ITU-R P.46-3 [2] recommendation addressed is available to predict a multi-point environment in a range of frequencies from 3 MHz up to 3, MHz, which means VHF and UHF bands, using adjustments through terrain clearance angle, effective height and many more parameters. Motivated by an optimization of the Land Mobile Service Private Networks, the ICP-ANACOM, National Authority of Communications facilitated the necessary material and people to provide an emission of 1 W, 4 dbm, in both bands in four kinds of environments. The environments chosen were: freespace; quasi-open, forest and sub-urban. After data processing and eliminating unwanted values [4], respecting the Lee criterion, the values obtained were analyzed and compared with those obtained from MONITORplus, a program that can provide a theoretical calculation of coverage area based in topographic maps with squares of 2 m, 1 m and in some special cases m by side resolution. Using this Manuscript received October X, 29. This work was supported by the ICP-ANACOM. Tiago A. A. Rodrigues, student of Integrated Master Degree in Electrical and Computer Engineering at Instituto Superior Técnico, Lisbon, Portugal. (Phone: 31 91 6949246; e-mail: de.azevedo.rodrigues@gmail.com) António J. C. B. Rodrigues is with the Electrical Engineering Department, Instituto Superior Técnico, Lisbon, Portugal. (e-mail: email@email.com). program some ITU-R P.46-3 parameters can be changed and selected to approach the theoretical values with the experimental ones. For example, terrain clearance angle in one or both antennas, climatic zone and clutter. Finally there is a discussion about the average difference between theoretical values and practical ones, represented by ε in db and their standard deviation, σ error in db, as well as a critique about the path loss exponent present in each environment. II. CASE STUDY A. Frequency band, modulation and bandwidth As previously mentioned, the frequency bands chosen were VHF and UHF because it is in those bands that the Private Networks from the Land Mobile service work. Thus, in this bands the frequencies selected were 17 MHz and 49.3 MHz respectively for VHF and UHF with a bandwidth B T =12 khz in a frequency modulation system. B. Environment and locations selected The environments chosen as explained before were: free space; quasi-open; forest and sub-urban. For free-space, quasiopen and forest the measures were taken in Catapereiro grange where the transmission antenna was fixed on h elev =23 m over mean sea level in VHF and UHF emission. On the other hand, for sub-urban environment two different locations were chosen, Tercena and Queijas where the broadcast antenna was fixed on h elev = 121 m and 133 m respectively. In all previous situations, the antenna reception was fixed at 2.8 m on a mobile van. C. Transmission set-up To obtain an EIRP = 4 dbm the following set-up circuit, figure 1, was taken into account. Figure 1: Transmission set-up

2 In short, with the target 4 dbm in the output of the transmission antenna, the generator must feed the set-up circuit with -6.22 dbm and -3.6 dbm respectively for VHF and UHF frequencies in study, considering an antenna emission gain of 2 dbi and an amplifier gain varying from 43.97 db for f = 17 MHz to 4. db in the case of 49.3 MHz. III. DATA PROCESSING Knowing, that a good data processing leads to a good conclusion [4], the exported data from the van program must be carefully managed. In this case, when the data was exported using Lee criterion provided by the program inside the van, the next step was to just import it in MONITORplus that was able to transform the values of dbµv, in electric field given in dbµv/m using the following equation. E ( f ) = V + 2 log - 29. - [ dbmv / m] [ dbmv ] [ MHz ] E / - G ant + L - G [ dbi ] [ db] ant[ dbi ] Where [ dbm V m] represents the field obtained from measured values V[ db m V ], accounting the losses, L [ db ], from cables and gains from antenna reception, G. ant [ dbi ] On the other hand, if the values were exported as raw data, after importing by MONITORplus, just to transform from tension to field respecting equation (1), they were again exported in.xml and imported in Matlab where the values were chosen in a way to respect Lee criterion and minimize the fast fading through a program previously built for this purpose. A. Lee criterion and fast fading Fast fading, also known as Rayleigh fading, represents the influence of diffraction and reflection in the surrounding environment near the reception antenna [8][9]. To avoid the fast fading, Lee proposed a theory where an average must be performed with a length of 2L=4λ. The samples, in this situation, must be separated by d sample =.8λ from each other, to guarantee independent samples. In theory, this means N= in 2L length to reach a confidence interval of 1 db [][6]. Thus, in this case study, accounting for the used frequencies, it leads. Table 1: Distance [m] on d sample and 2L Distance [m] d sample 2L 17 1.41 7.9 Frequency [MHz] 49.3.2 26.13 However, as expected it is almost impossible to have N= as the number of samples. Normally, Lee says that N=36 is the minimum value to get a confidence interval of 1.17 db, but for the present paper, respecting the following equation, the number of samples for each 4λ can be seen in table 2 as well as the confidence interval []. (1) ö ç æ.862 IC = 2 log 1 + (2) è N ø Table 2: N and IC for each study case. f [MHz] N IC [db] Free space 17 41.1 1.1 49.3 33.3 1.21 Quasi-open 17 38.3 1.13 49.3 34. 1.2 Forest 17 44.9 1. 49.3 34. 1.19 Queijas 17 42.2 1.8 49.3 3. 1.18 Tercena 17 41.7 1.9 49.3 32.8 1.22 IV. DATA ANALYSIS A good congruence between simulations supported on topographic maps with 2 m and 1 m side resolution was observed. Thus, from now, the analysis will be based on 2 m side resolution. The 1 m can be useful in case the investigated areas are bigger, saving computer processing time. At m side resolution, there are maps merely for some places of the biggest cities in Portugal, not for example in rural areas. A. Free space environment In the following table 3, the results obtained from the average difference between measured values and simulated values, ε [db], and its standard deviation σ error [db] will be presented. The simulated values were based in free-space recommendation from ITU provided by MONITORplus [3]. In figures 2 and 3 the evolution of electric field in relation with the broadcast antenna distance can be observed, as well as its difference, ΔE [db]. Table 3: Free space resolution 2 m f [MHz] Lee using program Recommendation Free - space Statistic ε [db] σ error [db] 17-9.44 2.48 49.3-4.4 3.38 Lee exported by Van Recommendation Free - space Statistic ε [db] σ error [db] 17-9. 2.6 49.3-4.2 3.48

3 1 9 9 8 8 7 7 Measured 6 6.1.2.3.4..6.7.8.9 1 Figure 2: Electrical field measured and simulated. Table 4: Quasi open resolution 2 m. f [MHz] Lee using program 17 9.23 1.86 9.2 1.87 49.3.11 3.61.7 3.61 Lee exported by Van 17 9.29 1.93 9.27 1.93 49.3.29 3.1.2 3.1 After table 4, figures 4 and represent the behavior of the electric field and the difference along the broadcast antenna distance for the frequency present in VHF band. 8 8 7 DE [db] 1 7 6 6 Measured - 4-1 -.1.2.3.4..6.7.8.9 1 Figure 3: Difference between measured and simulations. 4 1 1. 2 2. 3 3. 4 Figure 4: Electrical field measured and simulated. 1 As it can be seen, through the previous figures 2, 3 and table 3 the curve that fits the measured values better is the simulation using the free space recommendation, although in the case of VHF the simulation has an optimistic behavior of -9.44 db with a standard deviation of 2.48. On the other hand, the UHF has an error of -4.4 db and 3.38 db for standard deviation, using the Matlab program to process data. The values obtained by ITU-R P.46-3 simulations with and without TCA are not displayed here, due to the difference in values being greater. B. Quasi-open environment In this environment the values obtained through ITU-R P.46-3 simulations and measured values were analyzed. DE [db] - -1 - -2-2 -3 1 1. 2 2. 3 3. 4 Figure : Difference between measured and simulations. Looking to the figures 4, and table 4 one can detect that using the recommendation ITU-R P.46-3 with TCA flag active in both antennas can reach the best results, although the difference still is greater than wished, but now in a pessimistic

4 way. It means that the value measured is greater than the simulated using MONITORplus with ITU-R P.46-3 recommendation. The use of TCA flag also in the reception antenna does not bring too many advantages relative to the use of it just in the emission antenna as was done for all simulations. Worth pointing out that for UHF, the average error is greater than in the VHF band, but the same does not happen using free space recommendation. DE [db] 1 - -1 - C. Forest environment For this last environment measured in Catapereiro grange the comparison was made with ITU-R P.46-3 activating as well the TCA flag. Thus, leads Table : Forest resolution 2 m. f [MHz] Lee using program 17 9.2 2.1 8.99 2.1 49.3 12.4 3.8 12. 3.8 Lee exported by Van 17 9.8 1.9 9. 1.9 49.3 12.17 3.77 12.13 3.77 Once again, the figure 6 shows the relation between electric field along the broadcast antenna distance and the figure 7 shows the evolution of electric field difference of values with the distance merely for f = 17 MHz. 7 7 6 6 4 4 Measured 3 3.6 3.8 4 4.2 4.4 4.6 4.8 Figure 6: Electrical field measured and simulated. -2-2 -3-3 3.6 3.8 4 4.2 4.4 4.6 4.8 Figure 7: Difference between measured and simulations. Analyzing the evolution of the average error and the standard deviation since the previous environment a positive evolution of the average error can be detected, which means that the error from quasi open to forest environments decreases, however the σ err [db] has a small increase, but not comparable with the decrease of the error. In short, for this environment there is a better approximation than the others previously approached, in the best case, an error of 8.99 db with a standard deviation of 2.1 db for VHF in study is obtained and 12. db of error accounting with 3.8 db of standard deviation in UHF band in study. D. Sub-urban environment Finally the last case in study is the sub-urban. After an evaluation made previously, from this point, the values respecting Lee criterion exported directly by the van were not taken into account, due to a good congruence between the ones exported and the values obtained from raw data after processing by the built program. To facilitate, the two locals are presented together in a way to get the best conclusions. Thus, the next table 6 shows the values obtained from the difference between practical and theoretical study. Table 6: Sub-urban resolution 2 m. f [MHz] Lee using program Queijas 17 -.91 4.43 4.82 4.14 Tercena 17-6.6.71 3.3.78 Queijas 49.3-4.8.73 2.84.14 Tercena 49.3 -.39 6.12 6.42.76 As an example, the following figures 8 and 9 represent the evolution of the electric field as well as the difference between practical values and theoretical for Queijas in f = 49.3 MHz, UHF band.

DE[dB] 1 9 9 8 8 7 7 6 6 Measured.2.3.4..6.7.8.9 1 1.1 1.2 3 2 2 1 - -1 - Figure 8: Electrical field measured and simulated. -2.2.3.4..6.7.8.9 1 1.1 1.2 Figure 9: Difference between measured and simulations. Looking to the previous table 6 and figures 8 and 9, the suburban environment yields the best result which means the average difference is the smallest. Error values between -.91 db and -6.6 db can be seen with the standard deviation suffering a small increase, where those values can reach from 4.43 db to 6.12 db in an optimistic behavior, using only TCA for transmitter antenna (ITU 46-3). The use of TCA correction in both antennas (ITU 46-3 + TCA) appears with a pessimistic behavior meaning that measured values were greater than simulated ones, is this case from 2.84 db to 6.42 db. Comparing with the previous environments, the influence of TCA flag is greater due to the topography of the terrain. Observing figure 8, it can be seen that the simulated values using TCA flag for both antennas fits the shape of the measured values better. Path loss exponent Knowing that path loss exponent (n) represents the variation of radio-frequency, RF, signal along the route and it has a great sensitivity in each scenario, a short discussion is presented here about the obtained values. Considering the following equation, leads [8][9]. P r ( d ) = P ( ) - ç r d 1 n log è d ø Where d represents the initial distance, in this paper it represents the distance from the first sample to the emitter antenna. After some algebric manipulation to get equation (3) in a linear regression form and obtain the slope, the path-loss exponent obtained is listed in table 7: Table 7: Path loss exponent for each environment f [MHz] n Free space 17 3.2 49.3 3.7 Quasi-open 17 4.8 49.3 4.23 Forest 17 4.89 49.3 6.8 Queijas 17 2.88 49.3 1.9 Tercena 17 2.67 49.3 2. Analyzing table 7 some unexpected values can be seen, at least for the sub-urban environment as well as for forest in UHF band. These values can be justified through factors like: the average height from sea level in sub-urban is 133 m for Tercena and 121 m for Queija whereas just 23 m for the rest of the environments; for the first three environments the topography is almost flat, the same does not happen to suburban places; the distance when values start being recorded are different for all environments and finally in sub-urban locations there are is the possibility of the formation of street canyon models [1]. V. CONCLUSION In short, considering that MONITORplus respects the recommendation ITU-R P.46-3 the method used in this paper, achieves the best results for sub-urban environments with an optimistic or a pessimistic approach depending whether TCA is used merely for one antenna or for both. However the use of TCA flag (ITU 46-3 + TCA) fits best in the shape of measured samples with an average difference between 2.84 db to 6.42 db with a standard deviation form 4.14 db and.76 db using 2 m side resolution. Despite the results not being too different, when possible, m side resolution should be used, because it represents the topographic behavior of terrain in the best way. ACKNOWLEDGEMENT Tiago A. A. Rodrigues thanks to those which contributed for this project, particularly to professor António Rodrigues, the æ d ö (3)

6 ICP-ANACOM members Miguel Capela and José Gama. Final acknowledgement goes to António Eira, which helped with the English version of the present paper. REFERENCES [1] Laszlo Solymar, Getting the Message: A history of communications, Oxford University Press, 1999. [2] International Communications Union, ITU-R P.46-3 Method for point-to-area for terrestrial services in frequency ranges 3 MHz to 3 MHz, ITU Recommendation, 27. [3] International Communications Union, ITU-R PN.2-2 Calculation of free-space attenuation, ITU Recommendation, 1994. [4] International Communications Union, ITU-R SM.178 Fieldstrength measurements along a route with geographical coordinate registrations, ITU Recommendation, 2. [] William C.Y. Lee, Mobile communications design fundamentals, Wiley Series in Telecommunications, second edition, 1993. [6] William C.Y. Lee, Estimate of Local Average Power of a Mobile Radio Signal, IEEE Transactions on Vehicular Technology, vol. VT-34, pp. 22 27, Feb. 198. [7] Erik Östlin, Hajime Suzuki and Hans-Jürgen Zepernick, Evaluation of the propagation model Recommendation ITU-R P.46 for Mobile Services in Rural Australia IEEE Transactions on Vehicular Technology, vol. VT-7, pp. 38-1, 28. [8] David Parsons, The mobile radio propagation channel, Pentech Press, 1992. [9] T. S. Rappaport, Wireless Communications Principles and Practice, Prentice Hall, 22. [1] A. A. Zalava, S R. Saunders, Antennas and propagation for wireless communications systems, John Wiley & Sons, second edition, 27. [11] D. C. Montgomery, G. C. Runger, Applied statistics and probability for Engineers John Wiley & Sons, third edition, 23.